How to create a judging rubric for your competition

Published July 14, 2026 · Updated July 14, 2026 · By the Continuous Cup team

A judging rubric is the set of criteria your judges score every entry against, each on a defined scale, each weighted by how much it should count toward the final result. To create one: decide what excellence means for this competition, choose four to eight criteria that measure it, pick a scale and stick to it, write descriptors so a "7" means the same thing to every judge, weight the criteria to reflect what the competition is really about, and pilot the whole thing before the event.

The order matters. A rubric with the right criteria but a vague scale still produces scores judges cannot defend, and a well-scaled rubric with the wrong criteria still measures the wrong thing accurately. This guide walks through each step, then works through a full example rubric for a baking contest.

Start from what excellence means for this competition

Before you list a single criterion, write one or two plain sentences describing what a winning entry actually looks like in your competition. For a chili cook-off, is the point balance, or heat? For a race, is a clean run worth more than a fast one? Every criterion you choose should trace back to that sentence. Skip this step and you end up borrowing a generic rubric built for someone else's contest, which is why downloaded templates so often feel slightly off no matter how you tweak the numbers.

Choose four to eight criteria

Fewer criteria than you think you need is usually the right call. Each one should be independently observable: a judge should be able to score it without first deciding how they feel about the entry overall. Watch for overlap, "presentation" and "creativity" often reward the same thing twice, and so do "aroma" and "flavor" in a tasting rubric. A quick test: if two judges could disagree about one criterion without disagreeing about a second, the two are measuring different things and both belong. If they always move together, drop one.

Pick a scale and stick to it

Compare a 1 to 5 scale, a 1 to 10 scale, and a 0 to 100 scale, and pick one for the whole rubric rather than mixing them across criteria. A 1 to 5 scale is fast but coarse, two genuinely different entries can land on the same number with nowhere to go. A 1 to 10 scale gives enough granularity to separate entries without asking judges to distinguish a 71 from a 74, a call almost nobody can make consistently. A 0 to 100 scale looks precise, but in practice most untrained judges compress their scores into a narrow band anyway, so the extra range does not buy much. Whatever you pick, expect judges to avoid the true endpoints unless you give them a reason to use them, which is what level descriptors are for.

Write level descriptors so two judges read a score the same way

A number alone does not tell a judge what it means. For each criterion, describe in plain, observable terms what separates a low score from a middle score from a high score, tied to things a judge can see, taste, or measure, not adjectives alone. "A 9 or 10 has no visible flaws in shape or finish" tells a judge something to check for. "A 9 or 10 is excellent" does not. Descriptors are also what let you defend a result afterward: if an entrant asks why they scored a 6 and not an 8, the descriptor is the answer, not the judge's memory.

Weight the criteria

Weights encode what the competition is really about, so set them deliberately rather than splitting evenly out of habit. If flavor matters more than presentation, flavor should carry more weight, in numbers, not just in how judges happen to feel. Show the math plainly: a weighted score is the sum of each criterion's score multiplied by its weight. Scoring Appearance 8, Texture 7, Flavor 9, and Creativity 6 on a 1 to 10 scale, with weights of 15%, 25%, 40%, and 20%, gives (8 x 0.15) + (7 x 0.25) + (9 x 0.40) + (6 x 0.20), which comes to 7.75. Nothing about that math should be a mystery to anyone who asks how a result was reached.

Pilot the rubric on a few sample entries before the event

Hand the draft rubric to two or three people and have them score a handful of entries, past entries if you have them, sample products if you do not, before you commit to it for the real event. Watch for the places they hesitate, the criteria they interpret differently, and entries that tie when they should not. A rubric that reads cleanly on paper often falls apart the first time a real person applies it to a real entry, and it is far cheaper to find that out in a pilot than mid-competition.

Calibrate the panel

Before judging opens, walk the whole panel through one shared anchor entry together and have everyone score it independently, then compare. The point is not to force agreement, it is to surface where judges are reading the same descriptor differently, so you catch the gap before it shows up across fifty scored entries instead of one. A calibrated panel produces scores that are more consistent with each other even when judges still disagree on individual entries, which is the realistic goal, not perfect unanimity.

Common rubric mistakes

A few mistakes account for most of the rubrics that cause trouble later:

• Vague criteria. A criterion called "quality" is a placeholder for whatever the judge already thinks. Replace it with something specific enough that two judges would look at the same features.
• Too many criteria. Past eight or so, judges stop weighing each one and start scoring on overall impression anyway, which defeats the point of having a rubric.
• Hidden double weighting. Two criteria that measure the same underlying quality quietly count it twice, even though the rubric looks balanced on paper.
• Scales judges use unevenly. If one judge treats a 7 as average and another treats it as excellent, the raw scores are not comparable, no matter how carefully the criteria were written. See how to score a competition fairly for how aggregation methods handle this.
• Changing the rubric mid-competition. Entries judged before and after the change are no longer measured the same way, and there is no clean way to reconcile the two.

A sample rubric: judging a baking contest

A full example, scored on a 1 to 10 scale per criterion, with weights reflecting that flavor matters most:

A judge scoring one entry Appearance 8, Texture 7, Flavor 9, and Creativity 6 produces a weighted score of 7.75, using the same math shown above. Every judge scores against the same four criteria, the same scale, and the same weights, so a 7.75 from one judge and a 7.75 from another describe entries the panel actually agrees are comparable.

How Continuous Cup applies the rubric once you have built it

Once your criteria, scale, and weights are settled, Continuous Cup stores them as a single rubric attached to the competition, not a spreadsheet formula someone has to maintain by hand. Every judge sees the identical scoring form on whatever device they are using, phone, tablet, or laptop, so there is no version drift between one judge's copy of the rubric and another's. Weights are applied automatically as scores come in, which removes the class of error where a formula range does not update or a weight gets typed wrong. For the rest of the event lifecycle, see how to run a judged competition and this overview of a judged-competition platform.

Building the rubric is only half the job. Combining multiple judges' scores into one result, trimming outliers, and handling a judge who scores harsher than the panel is covered separately in how to score a competition fairly. Moving off a spreadsheet? This comparison walks through what changes.